水电站机电技术2026,Vol.49Issue(10):15-19,5.DOI:10.13599/j.cnki.11-5130.2026.10.004
基于轴心轨迹图像的水电机组智能预警方法研究
Research on intelligent early warning methods for hydropower generators based on shaft centerline orbit images
章立 1陈涛 2李俊 1安有恒 1杜承志2
作者信息
- 1. 贵州乌江水电开发有限责任公司东风发电厂,贵州 贵阳 550000
- 2. 北京华科同安监控技术有限公司,北京 100041
- 折叠
摘要
Abstract
Different shaft centerline orbit images can reflect various operating states of hydropower generators.This paper proposes an intelligent early warning method for hydropower generators based on shaft centerline orbit images from the perspective of quantifying the morphological similarities and differences of shaft centerline orbits.Firstly,a database of normal shaft centerline orbit images of the hydroelectric generators is constructed according to the operating conditions and measuring point positions.Then,the singular value decomposition is adopted for noise reduction,and the Hu invariant moment features of the denoised images are extracted.The Euclidean distance between the invariant moment vectors is used as an intelligent early warning indicator to evaluate the health states of the hydropower generators.On this basis,by comparing the deviation between the real-time calculated early warning indicator value and the corresponding early warning indicator value in the database,intelligent state early warning for hydropower generators is realized based on the Pauta criterion.The actual operating data results of the hydropower generators demonstrate that the proposed method can effectively quantify the differences among different shaft centerline orbit images,exhibits higher sensitivity to early weak morphological changes,and holds the potential for early fault warning.关键词
水电机组/轴心轨迹/奇异值分解/Hu 不变矩Key words
hydropowerc generator/shaft centerline orbit/singular value decomposition/Hu invariant moment分类
信息技术与安全科学引用本文复制引用
章立,陈涛,李俊,安有恒,杜承志..基于轴心轨迹图像的水电机组智能预警方法研究[J].水电站机电技术,2026,49(10):15-19,5.